Review: Sony SCD-XA5400ES CD Player

I saw a 5400 on the workbench at Allen Wright's (Vacuum State Electronics) studio in Schaffhausen the other Saturday. It had the Terra Firma bit clock fitted, and Allen or his technician were going to fit the Level 7 VSE modification board the next week.

As you know, the 5400 is a totally different chip architecture to what Allen's mods worked with previously. Interested to know what changes were necessary to make it work.

I hear that it's done now and I hope to get a chance to hear the VSE modified 5400 before it returns to its owner in the US.

The proof of the pudding is in the eating, but it will be great if this well-priced unit turns out to be a nice platform for modifications.

Regards,

Cyrilmartin, I have Maggie 3.6R's now so I think 40khz + is enough extension to hear what's going on up top on the Sony! I am Mr. Mom...no company affiliation. Attack would be wrong, I am just very impressed with the 5400 and like to correct any misunderstandings regarding it's performance characteristics. As for breakin? Well, if you don't think digital needs a long time to burn in and settle down then you are sadly mistaken:O(

Alright, I did some digging and have a bit of information to offer to the general public!

The DAC used in this unit is a DSD1796 (Burr Brown / TI) - The "Triple SADAC" topology used in the 9000es is gone, likely because it isn't very cost effective, and unless properly time-shifted "à la Accuphase", offers limited benefits; it does reduce the noise floor somewhat, but getting lower than -115dB FS seems ludicrous. The DAC is followed by a OPA2132 FET input dual opamp per channel as I/V converter, followed by one JRC2114/channel for buffering the balanced outputs, and again an OPA2132 for the unbalanced one. This selection of opamps clearly highlight the midrange ambition of this player as there are many well regarding (and much highly priced) opamps to choose from. Good news for us: there will be a lot of possible mods available in the form of opamp replacement and power supply improvements! The headphone output is driven through a JRC2043 dual opamp- but, and this to my surprise, the volume control is inserted in series between the output of the opamp and the headphones- this doesn't guarantee performance at all output levels as the impedance (and by association the damping factor) of the headphone output will change with the level. Apparently the quality of that section was considered "secondary" and not worth spending money on...

The decoder IC is a CXD9927R from Sony- there isn't much available regarding this particular chipset, but my guess is that it is Sony's newer integrated DVD/SACD chipset (there is even a CVBS composite video output onboard). It is clocked by at 27MHz, another clue that this IC is dedicated to video as well. RAM and Flash are attached to it for system software storage and operations. This chipset supports the HDMI output natively, but also the S/Pdiff output, and the laser head. There is also an (unbranded) microcontroler, with a cute little connector that goes to the back of the device, likely for reprogramming. It is accessible through a small hole on the back panel, on top of the right balanced audio outputs. Both SACD and PCM signals are going to the DAC as expected, seemingly indicating that SACD is supported natively without downsampling.

The power supply is composed of 3 transformers, a small laminated unit for the control circuit, and two 0 core ones for the analog and digital sections respectively.0 cores have the primary and secondary windings located "side by side", reducing high frequency coupling, at the cost of a lower efficiency. They seem however oversized and do not get too hot during operation. The supply is classically built on 78xx/79xx series regulators with relatively large heat sinks. No frills there! They are located on a separate board on the right side of the transport.

The control panel located behind the bezel is completely passive except for the display control. The jog button is a cost down version inherited from their low-end DVD players, and isn't very practical because 1) the steps are a tad too smooth and 2) the system software response is a tad slow. However, most of us will use the remote, making the point moot... The transport is a classic DVD drive similar to the ones that can be found in Sony DVD players. The chassis has a couple of beams to support the transport and make the player more rigid. There is also a small "shelve" to support both analog and digital boards. Manufacturing techniques are pretty standard for an audio product, although careful (through hole resistors in the audio section, sealed film caps for decoupling, surface mount film from Panasonic...).

All in all, this is a well designed unit where money was spent in the right places, leading to an accessible price tag. I must admit that it sounds much better than the sum of its parts- my compliments to the Sony engineers!

Update 2: I obtained the design white paper from Sony regarding the XA5400ES... As a side note, this paper should have been peer-reviewed as it contains many contradictory statements, and hints to the player having multichannel analog outputs, which it doesn't. However, it clearly states that CD upsampling is performed by the DAC IC, the DSD1796 from Burr Brown. The datasheet for this IC is available here.
As one can see in the datasheet page 8, the CD upsampling filter is a "regular" half band FIR filter; recent developments in digital filtering points to the pre-echo of that type of filter to being detrimental to sound quality, in particular sibilance (countless papers available on the AES). There are several alternative approaches to fix this issue: one of them is to use IIR filters instead of FIR, at the cost of a lot of phase distortion. Another approach is to use a "minimum phase" transform FIR filter, which translates the traditional symmetrical impulse response to asymmetrical at the cost of some phase distortion. This enables the designer to trade off pre-echos and phase to the optimal blend. The results are quite evocative, with better high-frequency extension and integration, especially when vocals are recorded with a large amount of high-frequency content. This filtering technique is in use in the Wolfson WM8741 DAC, but also in some outboard DACs such as the Cambridge Audio MagicDAC II (in conjuction with Anagram Technology Q5 upsampler), where the user can select his favorite upsampling method.

The SACD section of the player does also use a symmetrical FIR filter, but the very smooth LPF response required for SACD filtering enables the use of a very short sequence with limited pre-echo duration (see p. 39 and 40). This would explain why the 5400ES is so good at playing back SACD, but CD playback shows a slight hint of high frequency grunge. It is small though, and will be noticeable only compared to the best players out there.

All in all, it is a very good player. You would have to spend much more to get to this level of quality, and this is true for both CD and SACD playback. But it is possible to find a better CD player out there- it'll cost you though!

Wow, good stuff Cyrilmartin...thanks:O)